2 kernels and nvidia drivers installed at once and can there be 2 separate logins?
SUSE / openSUSEThis Forum is for the discussion of Suse Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
2 kernels and nvidia drivers installed at once and can there be 2 separate logins?
Hi!
I think this place could answer my questions. Keep in mind that SUSE and KDE (3.5) are new to me!!
I installed a dedicated realtime kernel for audio production (JAD2) according to instructions: I still have the openSUSE 11.0 default kernel.pae checked as installed in Yast software as well as the rt-kernel.
I want to keep 2 separate entries if possible in my grub page; one with the (j.engleh) rt-kernel, the other with the kernel.pae/ default, sort of as a backup and alternative log-in, which seem to I have- but with some problems!!
When I installed the rt-kernel, was I supposed to de-install the kernel-pae in Yast software?
Also how do you know for sure if your nvidia is ancient or legacy? I see that my Quadro4 900XGL is considered legacy at nvidia site so I assumed it to be same with j.engleh's repo! I have both nvidia-legacy-gfx and nvidia-legacy-gfx-kmp-rt installed. Should I remove the non -rt driver ?
Is it wrong to have two drivers/kernels installed at the same time or does openSUSE know and chooses what to use?
There was instruction to avoid updating the kernel not to loose the rt support- but I cannot use the auto-updater that keeps showing up in my bottom panel- it does not work- so I updated directly in Yast and did not update kernel,as instructed; (Yast actually showed me the dependancy conflict and offered me the selection). What do I do about the auto-updater in my panel that still shows me the same? 36 updates to install? Is there a way to have an-auto updater work for me like in Ubuntu? Or does my specialised kernel make it not workable?
Sorry I am plagued by many questions, thought I'd get it out of the way:
Can you work on both logins as separate SUSE entities (one is plain default SUSE 11.0, other is realtime?). I have a feeling that the answer is no! since something happened to my default login- I cannot start the Xserver in the pae login- even though I am at runlevel 5, it says "could not find the nvidia drivers" and thus "failed to start". I am afraid of doing installation repairs and removing my rt kernel- so help and advice is much needed (can I fix it in the CLI of my pae login)?
You can have many kernels installed on the same machine. Currently I have two. Each one has its own modules and drivers, because they are different versions. When I start one kernel, it loads its own modules, and when I start the other, it starts its own.
The kernel and the place where you run programs/work on the computer are divided. One is commonly called kernelspace, and the other userspace. So each kernel has its own kernelspace when it is running, but I am only using one userspace, or in other words, one operating environment. No matter which kernel I load, I can access the same files, have the same bookmarks on the internet, etc, because that is independent of the kernel.
Another way to look at it is like a car. The kernel is the engine, and tells the car wheels to go faster. The gas pedal (userspace) tells the engine (kernel) to go faster. If I switch engines (kernels), it is the same gas pedal (userspace) that tells the engine what to do.
Grub tells the computer which kernel to use, but your programs, users, etc, stay the same.
Usually if you have a specialized kernel, your distribution cannot update it, because it is not in its software repository, or database, so it ignores it. Any other programs which need to be updated or installed can usually be installed without trouble. This would be like updating your gas pedal.
Nvidia drivers must be installed for each kernel. Make sure that each kernel has its own Nvidia module, and that when you install one it doesn't install over the top of the other.
Thanks bcwagne for the explanation, I think I got it, exept for the software that is listed as "installed" in Yast. So you can actually see both kernels and graphic card drivers, no matter which kernel that you load from grub, right?
So the nvidia-legacy-gfx driver must be the wrong one for my default kernel right? Since I can't start the xserver! And the nvidia-legacy-gfx-kmp-rt must be o.k. since I can login fine with my specialized kernel. I think I will install the driver that comes straight from nvidia then and see if it works- it shouldn't overwrite the nv-rt driver since that is not from the official repository. And if I want to update my default kernel- I ask Yast to install and ignore the conficting specialized kernel rather than to de-install it of course!
Okay, so you should be able to boot with your regular kernel (not your rt kernel, since that one is working) and install the nvidia driver from nvidia. The nvidia installer SHOULD create a module just for that kernel, and leave your rt nvidia driver alone.
Note that you will have to have the kernel headers (part of the source for the kernel) installed for whichever kernel you are installing the drivers for. You will also have to have gcc and libc installed. I always forget that part until the nvidia installer complains at me.
I'm not sure how Yast handles kernels, but apt (on Debian) ignores anything which was not installed by the package manager. If your rt kernel was an rpm, you will have to tell Yast to ignore that one so it isn't changed at all.
Every time you update your kernel you will have to reinstall the nvidia driver, since it was built for a different version of the kernel. It's a bit of a pain, I know, but it's worth it. I told my package manager to not upgrade my kernel so I could just install the nvidia driver once and then not worry about it. So update if you want, then rebuild the driver, or don't update your kernel and don't rebuild the driver. Neither one is better, and neither one is really a big pain, so whatever suits you.
Good luck again...be sure to backup your rt kernel and modules first if you can, just in case it doesn't work the way you want.
You may want to use profiles if you need to load a different xorg.conf file due to using two different nvidia modules (if the name isn't "nvidia" in the Devices section for the realtime nvidia module.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.